Unbalanced Three Phase Load 6

[gcaudill]

I have an electrical system which consists of a delta-wye connected transformer supplying a delta connected load. One corner of the delta is grounded. I have the following load current measurements:
I leg1 = 1115A
I leg2 = 1000A
I leg3 = 1092A

Can someone tell me where the unbalanced current will go? Circulate in the load or exit to ground at the grounded corner? The ground is well connected, low impedance, and terminated to the main grounding bar at the switchgear.

 

[jbartos]

1.
Suggestion to previous posting: That is what the eng-tips are all about.
2.
Suggestion to gcaudill (Electrical) Apr 25, 2003 ///\\I am looking into getting phase angle measurements for the load. But as I said before, this is difficult, and somewhat dangerous.
///It just needs to sense the currents over conductors from X1, X2, and X3 transformer terminals to loads, respectively, e.g. by suitable probes, e.g. CTs and monitor the voltage on shunts by a three channel oscilloscope, and notice the angle shifts between currents Ix1, Ix2, and Ix3. The rest will be accurate and easy then.\\

 

[Shortstub]

Gentlemen, Gentlemen

How about a breather? Let's see if we can a least agree on the load configuration:

Gcaudill (please confirm) has stated that the delta-connected load has three legs which have been identified as leg 1, leg 2, and leg 3. For clarification I will call them segments A, B, and C, and the resultant delta arrangement follows:

o A segment has as terminal points sheave 2 and 3.
o B segment has as terminal points sheave 1 and 2.
o C segment has as terminal points sheave 3 and 4.

Now then, that means segments B and C are in series with their respective feeder cables, while segment A has no additional series impedance:

o B has a 10'x5/C feeder cable from its sheave 1 end to transformer terminal X2.
o C has a 35'x5/C feeder from its sheave 4 end to the same transformer terminal X2.
o A and B, have one of their common corners, connected to transformer terminal X1, via a 35'x5/C feeder.
o A and B, have their other common corner connected to transformer terminal X3, via a 15'x5/C feeder.

Gcaudill, please also confirm what the "leg" currents are. Are they measuring the delta-load segments? Or, are they measurements of the feeder conductors between the transformer and the sheaves to which they are connected?

Jghrist & Jbartos,
The 250MCM cable X/R ratio seems to be too high, unless you are somehow compensating for mutual reactance.

 

[Shortstub]

Correction to my last post:

o The common corner of segments A and B, are connected to transformer terminal X1, via a 35'x5/C feeder.
o The common corner of segments A and C, are connected to transformer terminal X3, via a 15'x5/C feeder.

 

[jghrist]

Shortstub,

You are right about the X/R ratio. I tried to guesstimate the reactance from NEC Table 9, but used ohm/km instead of ohm/1000' but did use ohm/1000' for resistance. The reactance is problematic because each phase conductor is a bundle of 10 (not 5 according to gcaudill's Apr 17 posting) cables. I calculate a reactance at one foot spacing of 0.0576 ohm/1000' for the bundle if it comprises three layers, 3/c bottom, 4/c middle, and 3/c top of 250 MCM RHH copper wire. If the three phase bundles were laying side by side, the reactance per phase would be 0.0315 ohm/1000'. If the equivalent spacing were two feet, the reactance would be 0.0735 ohm/1000'.

The resistance would be 1/10 of 0.052 ohm/1000' if not in metallic conduit.

We don't know the spacing, but we do know that the mutual reactances between phases are not equal, so things are a bit messy. One of the phase connections (X2) is actually two separate bundles going to different sheaves. With my corrected smaller reactance, I feel better about ignoring the unbalanced cable configuration. The reactance of a 35 foot length would be around one milliohm if the cables were lying side by side.

I'm not sure what you mean by "segments B and C are in series with their respective feeder cables, while segment A has no additional series impedance". The segments are connected between phases, not in series with the feeder cables. Your segment A is between X1 and X3, segment B is between X1 and X2, and segment C is between X2 and X3.

 

[Shortstub]

Jghrist, "one picture is worth a thousand words"! I can't give you the picture, so here are the words:

Call the three annealed-wire element impedances, Za, Zb, and Zc:

o Za, located between sheaves 2 and 3.
o Zb, located between sheaves 1 and 2.
o Zc, located between sheaves 2 and 4.

Identify the four feeder elements taken from the transformer terminals as Z1, Z2, Z3, and Z4:

o Z1, located between sheave 1 and X2.
o Z2, located between sheave 2 and X1.
o Z3, located between sheave 3 and X3.
o Z4, located between sheave 4 and X2.

Remember, Gcaudill said one corner of the delta load was groundED. He meant the corner was connected via X2, the intentionally grounded terminal of the transformer. Not the"corner" of the delta elements!

Thus, the three delta-load segment impedances have to be "adjusted" to include feeder impedances. They are, ZA, ZB, and ZC, where:

o ZA = Za.
o ZB = Zb + Z1.
o ZC = Zc + Z4.

The final impedance to consider is the effect of the quench system. I called this ZQ between sheave 4 and ground.

 

[jghrist]

Shortcut,

I think I see your picture. I modelled the circuit with each of the cables and wire elements separately instead of combining them, and then wrote four loop equations. One of the loops just accounted for the ground path and really could have been handled as a parallel path to Z4. My loops, in terms of your nomenclature are:
X0-X3-Z3-sheave 3-Za-sheave 2-Z2-X1-X0
X2-Z1-sheave 1-Zb-Sheave 2-Za-sheave 3-Zc-sheave 4-Z4-X2
X0-X2-Z1-sheave 1-Zb-sheave 2-Z2-X1-X0
X2-Z4-sheave 4-ZQ-X2

Any theories on why changing the number of commutator brushes on sheaves 1&3 doubles the ground current? See gcaudill's Apr 21 post. Makes me think that there is a significant contact resistance between the sheaves and the wire which is reduced by increasing the number of brushes.

 

[wggei]

I have read the threads and this seems to be getting a lot of interest.

I beleive you have noted that the controls for this annealer have SCR's. No one to date has acknowledged harmonics. If the load had as much as 2.5% current harmonic at 1000A in each of the three phases of the load, the harmonic components would total approx 75A. This current would not add to zero in the normal sense.

 

[Shortstub]

Gcaudill,
Are you sure that sheaves 1 and 4, are not somehow connected together? If they are truly connected only at X2, then the the two feeder impedances are in series with their respective annealed-wire segments!

Weggi,
Could you identify the response showing the the annealer is SCR controlled!

Jghrist,
You can reduce the configuration to just two loops by delta-wye conversion of the annealer load, and maintaining a delta configuration of the transformers wye terminal X1, X2, and X3.

 

[gcaudill]

Sheaves 1 & 4 are connected by 10 250mcm cables each back to X2 of the x-former. The only intentional connection to ground is at X2 of the x-former. Sheaves 1-4 are isolated from ground.

The voltage controller is a SCR type. I am looking into measuring the current in the grounded conductor connected X2 to see if the current (78A) is at 60Hz or at some harmonic.

 

[gcaudill]

Here are some new measurements. They are much worse than the last batch regarding load balance.

X-former primary amps
I ph.1 = 216A
I ph.2 = 224A
I ph.3 = 246A
Seems okay.

X-former secondary amps
Conductor Gang 1 (Ph. X2 to shwave 1) = 866A
Conductor Gang 2 (Ph. X1 to sheave 2) = 1300A
Conductor Gang 3 (Ph. X3 to sheave 3) = 1153A
Conductor Gang 4 (Ph. X2 to sheave 4) = 606A

X-former secondary volts
X1 to gnd - 56VAC
X2 to gnd - 0V
X3 to gnd - 48VAC
Seems like a problem?

Load amps (delta)
Currents:
I leg1 = 836A
I leg2 = 673A
I leg3 = 529A
Seems very unbalanced?

X2 grounding conductor is carying 46A

Comments?

 

[busbar]

Does the primary-current imbalance affect other processes or equipment operation? What portion of the total plant load is represented in the annealer 480V, ±250A 3ø current?

For the given dynamitic load, I don’t think the primary-current unbalance is that unacceptable. Whether the SCR firing is based on a constant current, constant voltage or constant power, it will still vary with the moving product. Even if you redesigned the controls to use an individual 3ø full-wave bridge rectifier {and maybe separate transformers if SCR gating is done on the primary side} with DC out for each segment of the annealing process, the currents may be less unbalanced, but there will still be fluttering/variations observed in primary current. The only fix at that point may be huge LC components, probably at a prohibitive price.

Maybe I have misunderstood, but there a no zero-sequence component in the circuit serving the annealer, and no staggering negative-sequence current either.

 

[gcaudill]

The unbalance on the transformer primary is not of concern to me as it is only a small part of the total plant load. The SCR voltage controller is on the primary side of the transformer. The low ground current tells me the isolation is good.

I am somewhat concerned by the differences in the secondary voltages measured to ground. These voltages are actually measured X1-X2 and X3-X2. X2-ground measured 0V.

 

[jghrist]

gcaudill, That's a big voltage difference (X1-X2 = 56V, X3-X2 = 48V). What is X1-X3? What are the primary ø-ø voltages? What is the transformer impedance? Is it possible to measure the current in the wire between sheave 4 and the quencher?

Shortstub, I decided not to do the star-delta transformation so that I could calculate directly the current in each transformer winding and show that it adds to zero. Either way is equivalent.

wggei, With the delta-connected load and a delta-ungrounded wye transformer, there is no zero-sequence path so third-order harmonics would be suppressed. As gcaudill noted, the SCR is on the primary, so any harmonic currents from the SCR would have to go thru the transformer to get to the load.

 

[jbartos]

Suggestion to gcaudill (Electrical) Apr 29, 2003 marked ///\\Here are some new measurements. They are much worse than the last batch regarding load balance.

X-former primary amps
I ph.1 = 216A
I ph.2 = 224A
I ph.3 = 246A
Seems okay.
///The seems o.k. however; they confirm that there is an unbalanced load downstream. It is a good observation.\\X-former secondary amps
Conductor Gang 1 (Ph. X2 to shwave 1) = 866A
///Conductor gang 1 and conductor gang 4 shall have combined currents, i.e. vectorially summed, to obtain Ix2 current. It would be good if you could provide the phasor angle of the combined currents in conductor gang 1 and gang 4, with reference to, for example, conductor gang 2.\\Conductor Gang 2 (Ph. X1 to sheave 2) = 1300A
///It would be good if you could provide the phasor angle with reference to for example conductor gang 3.\\Conductor Gang 3 (Ph. X3 to sheave 3) = 1153A
///It would be good if you could provide the phasor angle with reference to for example conductor gang 1.\\Conductor Gang 4 (Ph. X2 to sheave 4) = 606A
///Conductor gang 4 and conductor gang 1 shall have combined currents, i.e. vectorially summed, to obtain Ix2 current.\\X-former secondary volts
X1 to gnd - 56VAC
X2 to gnd - 0V
X3 to gnd - 48VAC
Seems like a problem?
///The voltage analysis could take place after the current analysis is done. Incidentally, the current issue is your original concern.\\Load amps (delta)
Currents:
I leg1 = 836A
I leg2 = 673A
I leg3 = 529A
Seems very unbalanced?
///Yes. It is nice to know for the purpose of your questions in the original posting.\\X2 grounding conductor is carrying 46A
///It is a ground current which leaks to the ground and uses ground path to complete its circuit according to Kirchhoff Laws. It is called common-mode current and it is carrying zero sequence current too. The rest of load unbalance is done over the negative sequence currents in the main conductors.\\Comments?
///It is nothing new to those who are analyzing asymmetries including fault current values caused by various faults in power distribution systems.\\

 

[jghrist]

jbartos,

I finally got around to reviewing your Stevenson reference drawings. All show ground connected to the neutral. Do you know of any references that show zero-sequence equivalents of corner grounded systems? Either transformer windings with an isolated X0 as in the present case or delta windings, with one phase grounded?

Symmetrical components are very suitable for fault analysis, but I have found that trying to analyze unbalanced load situations with symmetrical components is difficult. There are some equivalent diagrams in the Westinghouse T&D Manual (Fig. 21 in the 4th Edition) but only with two phase loads identical and the third different.

 

[gcaudill]

Still looking into getting phase angle measurements all the way around the secondary and load.

 

[gcaudill]

Also looking into the possibility that a bad SCR is causing a low voltage on one phase of the transformer.

 

[gcaudill]

Also looking into the possibility that a bad SCR is causing a low voltage on one phase of the transformer. Does anyone know of web site with photos of phase-to-phase voltage traces of normal SCR output and when one SCR is shorted. The SCR's are arranged in a three phase anti-parallel configuration. They are controlled by a gate circuit and are phased on/off to control voltage amplitude. I could use this to show the electrician what to look for when troubleshooting.

 

[jbartos]

Suggestion: The currently posted data lead to approximate solutions. Therefore, the problem is ill posed. To have a well-posed problem, the angles among currents or an equivalent (that would lead to the angles among currents determination) are needed. The corner grounded star or delta connections are rather special grounding approaches.

 

[Shortstub]

Gcaudill,

It is impossible to have a shorted SCR without other abnormal manifestations in the electrical system.

Have you measured the current in the annealed-wire segment leaving sheave 4 enroute to the quench tank?

Are you able to make voltage measurements between sheaves? As well as the voltage from each sheave to "ground?"